How to design a questionnaire

By Dr. Rafiq Muhammad, MD, PhD · Updated June 2026

A questionnaire is a measuring instrument, and a badly worded item measures the wrong thing just as surely as a miscalibrated scale. The goal: questions that every respondent reads the same way, answers honestly, and finishes. Two principles do most of the work — write neutral, single-idea questions, and keep it as short as the research questions allow.

Open vs closed questions

ClosedOpen
AnswerFixed options (scale, multiple choice)Respondent’s own words
DataQuick to analyse, comparableRich, but slow to code
Use forMost itemsA few targeted “why” questions

Most surveys are mostly closed questions (often Likert scales) with a few open ones where you genuinely need the respondent’s own framing.

The wording mistakes that invalidate data

Order, length, and piloting

Open with easy, engaging questions; put sensitive or demographic items near the end. Watch for order effects (an earlier question priming a later one). Above all, mind length: completion time drives drop-off, and a fatigued respondent gives careless answers. Cut any item that doesn’t map to a research question, then pilot the whole thing on a handful of people before launch — piloting catches confusing items nothing else will.

Length is the silent killer of response rates. The free Survey Length Estimator predicts completion time from your item count and types, so you can trim before drop-off costs you data.

Get the free Data Collection toolkit

A questionnaire-design checklist, a measurement worksheet, and a data-management-plan starter — from Research Design Simplified. We’ll email you the download link.

One email with your download, then occasional research tips. One-click unsubscribe, anytime. We never sell your data.

Get Research Design Simplified

Frequently asked questions

Open vs closed questions?

Closed have fixed options (quick to analyse); open are free text (richer, slower). Most surveys are mostly closed.

What is a leading question?

One that nudges toward an answer (“our excellent service”). It biases responses — keep wording neutral.

What is a double-barrelled question?

One that asks about two things at once (“price and quality”). Split it into two questions.

How long should it be?

As short as your research questions allow — length drives drop-off. Estimate completion time, cut, and pilot.

What is a Likert scale? → Open the Survey Length Estimator →